Method for correcting deviation of drilling hole for grouting and water plugging in karst area subway tunnel limestone section

Abstract

The karst area subway tunnel limestone section grouting water plugging drilling deviation correction method relates to the field of tunnel and underground engineering construction technology, and comprises the following steps: obtaining real-time deformation data of a drilling tool assembly in the drilling process, and detecting geological feature information of a front end of the drilling tool assembly; based on the real-time deformation data and the geological feature information, a preset drilling tool dynamics model is used to reversely identify parameters representing a geological mechanics environment in which the drilling tool assembly is located; based on the geological mechanics environment parameters, a model predictive control method is used to generate active guiding instructions for guiding a drilling direction. The present application uses the technology of combining a forward-looking acoustic transducer array with a model predictive control to realize the pre-perception of the front geological information and the forward-looking guidance of the drilling trajectory, and solves the technical problems of sudden emptying of the drill bit, sticking of the drill or damage of the drilling tool caused by the reaction lag.

Description

Technical Field

[0001] This invention relates to the field of tunnel and underground engineering construction technology, and in particular to a method for correcting deviation in drilling holes for grouting and water plugging in limestone sections of subway tunnels in karst areas. Background Technology

[0002] In urban subway tunnel construction, especially when traversing limestone strata in karst areas, grouting for water plugging is a crucial and challenging preliminary step. This step requires drilling multiple precise grouting holes along the designed contour ahead of the tunnel excavation face, and then injecting grout into these holes to reinforce the surrounding rock and seal off groundwater. The drilling quality of these grouting holes directly determines the effectiveness of subsequent grouting and the safety of tunnel construction.

[0003] Currently, directional drilling technology used for this type of borehole construction typically relies on measurement-while-drilling (MWD) or logging-while-drilling (LWD) systems installed behind the drill bit to obtain attitude information of the drill string, such as the apex angle and azimuth angle. The surface or downhole control system applies control forces via guiding mechanisms (such as push blocks or variable-angle joints) based on the deviation of the measured values ​​from the designed trajectory, aiming to pull the borehole trajectory back onto the predetermined path.

[0004] However, in the extremely complex and heterogeneous strata of karst areas, existing technologies have revealed a series of insurmountable drawbacks when applied: First, the control logic of existing technologies is essentially a delayed response. The measuring tools are typically placed several meters away from the drill bit, meaning that after the borehole trajectory deviates, drilling must continue for a certain distance before the measuring system detects it. When the drill bit suddenly encounters a cave or large fissure, the measuring system often only issues an alarm after the drill string has already entered a geologically abnormal zone. By then, correction is too late, easily leading to accidents such as drill bit falling or jamming. The control behavior is entirely passive.

[0005] Secondly, traditional correction methods rely on a very singular decision-making basis. They depend almost entirely on the geometric deviation between the actual drilling trajectory and the designed trajectory. The control system is completely blind to the root cause of this deviation—the changes in the mechanical properties of the rock mass in front of and around the drill bit. This lack of awareness of the physical environment results in correction actions that are often mechanical and rigid, potentially causing excessive pressure between the drill bit and the hard rock wall. This not only exacerbates the severe vibration of the drill string but may even apply force in the wrong direction, leading to correction failure.

[0006] Furthermore, the objective function of existing control strategies has limitations. The core objective of their control algorithms is to eliminate trajectory deviations as quickly as possible. To achieve this single objective, the system may output the maximum correction force regardless of the cost. It does not comprehensively consider how to achieve smoother and more energy-efficient drilling while ensuring accuracy, and lacks a globally optimal control concept.

[0007] Finally, many existing intelligent guidance systems rely on communication between the wellbore and the surface to make complex decisions. Data is uploaded to the surface via mud pulses, etc., and then analyzed by operators or surface computers before commands are issued. This communication link has an inherent, non-negligible time delay. In karst areas, where geological conditions change rapidly, control commands calculated based on data from minutes or even earlier will have already changed in terms of the geological environment upon reaching the downhole actuators, thus significantly reducing the effectiveness of the commands. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a method for correcting borehole deviation in limestone sections of subway tunnels in karst areas through grouting and water plugging. This method solves the problems of low drilling accuracy, easy damage to drilling tools, and low overall efficiency caused by the inability to predict geological changes ahead and the single and lagging control strategy in complex geological conditions such as karst areas.

[0009] To achieve the above objectives, the present invention provides a method for correcting deviation through grouting and water plugging drilling in limestone sections of subway tunnels in karst areas, comprising the following steps: S1: Obtain real-time deformation data of the drill string assembly during the drilling process, and detect the geological features of the front end of the drill string assembly.

[0010] First, real-time deformation data of the drill string assembly during drilling is acquired. This task is exemplarily accomplished by a densely distributed fiber optic strain sensor array deployed along the drill string body. These sensors can continuously measure the strain distribution of the drill string under torsional, bending, and other loads with extremely high frequency and spatial resolution. By processing this massive amount of strain data in real time, the downhole microprocessor unit can reconstruct the three-dimensional deformation curve, bending curvature, and torque state of the entire drill string assembly in the wellbore, providing accurate measured input for subsequent mechanical analysis.

[0011] Secondly, the system actively detects geological features at the front end of the drill bit assembly. This task is exemplarily performed by a miniaturized forward-looking acoustic transducer array mounted behind the drill bit. This array periodically emits focused acoustic beams into the rock mass in front of the drill bit and receives echo signals reflected from different geological interfaces (such as lithological change surfaces, fissures, and cavities) to obtain the acoustic impedance distribution of the rock mass ahead. By analyzing information such as the flight time, amplitude, and phase of the echo signals, the system can "see" the geological structure distribution within a range several meters ahead, generating a dynamically updated geological profile, thereby identifying cavities, fissures, or hard-soft interfaces and providing early warning of impending geological upheavals.

[0012] S2: Based on the real-time deformation data and the geological feature information, the parameters characterizing the geomechanical environment of the drill string assembly are identified in reverse through a preset drill string dynamics model.

[0013] The drill string assembly is equipped with a downhole microprocessor unit, which is configured to: based on acquired real-time deformation data and geological feature information, identify parameters characterizing the geomechanical environment of the drill string assembly through a pre-defined drill string dynamics model. Specifically, the downhole microprocessor unit first calls a pre-defined partial differential equation model that can accurately describe the dynamic behavior of the drill string assembly. This model comprehensively considers internal and external factors such as the drill string's elastic modulus, cross-sectional moment of inertia, linear density, and real-time drilling pressure. Subsequently, the system processes the three-dimensional deformation data measured in real-time in S1 (i.e., The known output is substituted into the dynamic model. Since the model includes forces representing rock reaction (…),… Since the term of the equation is unknown, the system can solve the equation in reverse to deduce a rock reaction force that causes the currently observed drill string deformation. Furthermore, based on rock mechanics theory... The system ultimately identified the core parameter that characterizes the reaction force—the geomechanical stiffness matrix. The elements of this matrix (such as...) (etc.) directly quantifies the rock mass's hardness, anisotropy, and coupling characteristics, thereby transforming the invisible subsurface environment into precise mathematical parameters that can be used for control calculations. The geological feature information obtained in S1 can be used to constrain and verify the results of this reverse identification, significantly improving the accuracy and robustness of the identification.

[0014] S3: Based on the aforementioned geomechanical environmental parameters, a model prediction control method is used to generate active guidance commands to guide the drilling direction.

[0015] The downhole microprocessor unit is also configured to generate active guidance commands based on the aforementioned geomechanical environment parameters using a model predictive control method to drive the active guidance actuator. Specifically, the MPC module within the downhole microprocessor unit uses the parameterized model identified in S2 as its internal virtual drilling system. Within this virtual system, it predicts how a series of different alternative control commands (such as applying different thrusts to different push blocks) will produce drilling trajectories and drill string states within a short prediction time domain (e.g., dozens of control cycles). The core of the MPC algorithm is to solve a multi-objective optimization problem, the goal of which is to find a set of parameters that satisfy a comprehensive cost function. The optimal control sequence is minimized. This cost function, exemplarily, includes several sub-items, each penalizing: the deviation between the predicted and designed trajectories, the magnitude of the unbalanced force on the drill bit (based on...). Matrix calculation), and rock-breaking energy per unit advance (also based on...) (Matrix-based rock hardness assessment). Global performance optimization is achieved by minimizing this comprehensive cost function.

[0016] S4: Drive the active guidance actuator in the drill string assembly to execute the active guidance command.

[0017] It features a controller that precisely adjusts the pressure flowing to the hydraulic cylinders of each push block via a high-precision servo control loop, driving the designated push block to extend and apply lateral thrust to the well wall in the required magnitude and direction. Each push block integrates a force sensor for real-time feedback of the actual output force, thus forming a force control closed loop to ensure execution accuracy.

[0018] This physical thrust alters the force state at the drill bit, guiding it to deflect in the desired direction. Crucially, the minute changes in drill string deformation and attitude caused by this physical action are immediately captured by the sensing system in S1, serving as the initial input for the next control cycle and initiating a new round of the "perception-identification-decision-execution" process. This high-frequency, rolling closed-loop process enables the invention to respond quickly, accurately, and intelligently to complex and ever-changing geological environments.

[0019] As a further improvement of the present invention, the geomechanical environmental parameters are a geomechanical stiffness matrix, and each element of the matrix represents the support stiffness and coupling effect of the rock mass around the drill bit in different directions.

[0020] As a further improvement of the present invention, in the reverse identification step, the geological feature information is used as a priori constraint and substituted into the reverse solution algorithm for solving the geomechanical stiffness matrix.

[0021] As a further improvement of the present invention, the cost function used in the model predictive control method includes at least one trajectory tracking term for penalizing the geometric deviation between the predicted trajectory and the design trajectory.

[0022] As a further improvement of the present invention, the cost function also includes a drill bit force balance term calculated based on the geomechanical stiffness matrix to penalize drill bit force imbalance.

[0023] As a further improvement of the present invention, the cost function also includes a rock-breaking energy term estimated based on the geomechanical stiffness matrix and drilling parameters, used to penalize high energy consumption.

[0024] This invention provides a method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas. It has the following beneficial effects: 1. This invention utilizes a combination of forward-looking acoustic transducer arrays and model predictive control (MPC) to achieve advance perception of geological information ahead and proactive guidance of the drilling trajectory. This method enables the drill bit to avoid or adaptively adjust before encountering adverse geological formations such as caves and fractures. Compared to the commonly used post-drilling measurement and passive correction schemes in existing technologies, this invention solves the technical problem that its delayed response often leads to sudden drill bit slippage, stuck drill bit, or drill bit damage.

[0025] 2. This invention creatively establishes a geomechanical stiffness matrix, parameterizing the invisible rock mass mechanical environment. By integrating real-time deformation data of the drill string itself with geological feature information detected at the front end, it enables the control system to understand the root cause of borehole deflection from a physical perspective. This differs from existing technologies that rely solely on feedback control based on geometric deviations of the trajectory. This invention overcomes the inherent defects of traditional control methods, such as blind control and susceptibility to drill string vibration due to a lack of understanding of geological formation.

[0026] 3. This invention employs a multi-objective cost function that integrates trajectory tracking accuracy, drill bit force balance, and rock-breaking efficiency for comprehensive optimization decision-making. This control strategy ensures drilling accuracy while also considering drill string stability and drilling economy, seeking the globally optimal drilling state. Compared to existing single-objective control methods that typically focus only on trajectory tracking, it overcomes the shortcomings of sacrificing drill string life and reducing drilling efficiency for forced correction.

[0027] 4. This invention highly integrates the entire process of perception, modeling, decision-making, and execution into a downhole microprocessor unit, forming a high-speed, autonomous closed-loop control system behind the drill bit. This ensures instantaneous response and processing capabilities to geological changes. It changes the existing technology's heavy reliance on ground monitoring centers and slow uplink and downlink data communication for intervention, fundamentally solving the problem of severe control lag caused by communication delays, and significantly improving the timeliness and reliability of drilling in complex strata with rapidly changing geological conditions, such as karst areas. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the method steps of the present invention; Figure 2 This is a schematic diagram of the drill string assembly of the present invention; Figure 3 This is a schematic diagram of the downhole microprocessor unit structure of the present invention; Figure 4 This is a schematic diagram of the active guiding actuator of the present invention; Figure 5This is a schematic diagram of the control closed loop of the present invention. Detailed Implementation

[0029] The present invention will be further described below with reference to specific embodiments and accompanying drawings.

[0030] like Figure 1-5 As shown in the figure, this invention provides a method for correcting deviation in drilling holes for grouting and water plugging in limestone sections of subway tunnels in karst areas, including the following steps: S1: Obtain real-time deformation data of the drill string assembly during the drilling process, and detect the geological features of the front end of the drill string assembly; Specifically, to implement the method described in this invention, a core device is connected behind the drill bit in the drill assembly (BHA). This device includes a sensing module, a downhole microprocessor unit, and an active guidance actuator for implementing the steps of this invention. The following will describe in detail the specific implementation process of acquiring real-time deformation data of the drill assembly and detecting front-end geological feature information in the method of this invention.

[0031] The steps of acquiring real-time deformation data of the drill string assembly during drilling and detecting geological features through the front end of the drill string assembly are completed collaboratively by the sensing module and the downhole microprocessor unit. The sensing module forms the physical basis for data acquisition, while the downhole microprocessor unit is responsible for receiving and processing the data. The two are electrically connected and transmit data through an internal high-speed data bus.

[0032] Specifically, the real-time deformation data of the drill string assembly is obtained through a distributed strain sensor array deployed on a specific drill collar of the drill string assembly. For example, the array consists of multiple optical fibers embedded with fiber Bragg grating (FBG) sensors. The optical fibers are precisely deployed in a spiral pattern on the inner and outer walls of the drill collar, and their spatial layout covers at least four circumferential directions: 0 degrees, 90 degrees, 180 degrees, and 270 degrees.

[0033] During drilling, the fiber Bragg grating demodulation module built into the downhole microprocessor unit emits broadband light to the FBG sensor at a preset high frequency (e.g., 1000 Hz) and monitors its reflection spectrum in real time. When the drill collar bends or twists due to force, the FBG sensors at various locations undergo axial stretching or compression, causing a shift in their central reflection wavelength. The demodulation module collects this wavelength shift. .

[0034] After receiving the wavelength drift, the downhole microprocessor unit first converts it into axial strain at the corresponding position based on the sensing principle of the fiber Bragg grating. .

[0035] Subsequently, the downhole microprocessor unit uses the strain-curvature relationship in beam theory to calculate the bending curvature of the cross section in two orthogonal planes (e.g., horizontal and vertical planes) based on the strain values ​​at different locations along the circumference of the drill collar. This relationship can be expressed by the following formula: ; In its formula, The drill collar in axial position and time The curvature of the bend, This refers to the axial strain at the corresponding position on the drill collar surface, calculated based on the wavelength drift. The radial distance from the neutral axis of the drill collar to the fiber core of the FBG sensor is a known geometric parameter.

[0036] After obtaining the discrete curvature distribution along the entire axis of the sensing drill collar, the downhole microprocessor unit uses a numerical integration algorithm to process the curvature function. Along the axial direction A first integration yields the slope (i.e., rotation angle) of the drill collar's deflection curve; a second integration reconstructs the real-time deformation curve of the entire sensor drill collar in three-dimensional space, which is defined as... This curve accurately describes the actual bending shape of the drill string at the current moment.

[0037] Simultaneously, or alternately with the above steps, the step of detecting geological feature information at the front end of the drill string assembly is performed. This step is achieved by an acoustic transducer array disposed at the front end of the drill string assembly, preferably at the front face of the sensing module. This array contains multiple piezoelectric ceramic ultrasonic transducers evenly distributed in a ring, with operating frequencies exemplarily between 500 kHz and 2 MHz.

[0038] The downhole microprocessor unit drives the transducer array to actively transmit controlled acoustic pulses towards the rock mass in front of the drill bit at a preset pulse repetition frequency (e.g., 100 Hz). The acoustic pulses propagate in the drilling fluid and reach the rock mass interface ahead, where some of the energy is reflected to form an echo signal, which is captured by the same transducer array or a dedicated receiving transducer.

[0039] The downhole microprocessor unit acquires and analyzes the complete acoustic echo waveform. By calculating the flight time from emission to reception of the echo signal, the distance to the geological discontinuity interface ahead can be determined. Furthermore, by analyzing the amplitude of the echo signal, the reflection coefficient of the sound wave at the medium interface is calculated according to the following formula. : ; In its formula, The sound wave reflection coefficient, The amplitude of the received echo signal. The amplitude of the original transmitted signal. Let be the acoustic impedance of the drilling fluid, and be a known or measurable parameter. The acoustic impedance of the rock mass to be probed is shown.

[0040] By solving the above equation, the downhole microprocessor unit can infer the acoustic impedance of the rock mass ahead. By scanning and integrating the detection results from all transducers, a qualitative acoustic impedance distribution map characterizing the area in front of the drill bit can be generated. This map can visually reveal geological features: areas with acoustic impedance values ​​close to zero indicate the presence of caves or large open fractures; areas with drastic step changes in acoustic impedance values ​​indicate the presence of interfaces between soft and hard rock layers.

[0041] Finally, the downhole microprocessor unit calculates the real-time deformation curve. The generated geological feature information (i.e., acoustic impedance distribution map) serves as two synchronous and crucial data input streams, providing data for the subsequent reverse identification step of the geomechanical environment. This process constitutes the data foundation of the intelligent control closed loop of this invention, providing timely and accurate real-world data for subsequent physical modeling, prediction, and decision-making.

[0042] S2: Based on the real-time deformation data and the geological feature information, the parameters characterizing the geomechanical environment of the drill bit assembly are identified in reverse through the preset drill bit dynamics model. The geomechanical environment parameters are a geomechanical stiffness matrix, and each element of the matrix characterizes the support stiffness and coupling effect of the rock mass around the drill bit in different directions.

[0043] Specifically, following the aforementioned steps, after the downhole microprocessor unit acquires the real-time deformation curve of the drill string assembly and the geological feature information of the front end, the method of this invention then performs reverse identification of geomechanical environmental parameters. This step is entirely completed within the downhole microprocessor unit by running a preset algorithm program, the core of which lies in establishing a physical model, parameterizing the geological environment, and performing reverse solving that incorporates prior information.

[0044] First, the downhole microprocessor unit loads a pre-defined drill string dynamics model that describes the mechanical behavior of the drill string assembly. For example, this model is an Euler-Bernoulli dynamic beam-column model considering the influence of axial pressure, mathematically expressed as a partial differential equation: ; In its formula, The elastic modulus of a drill collar reflects the stiffness properties of the material. The moment of inertia of a drill collar section represents the effect of the material's shape on its bending ability. This represents the fourth derivative of the drill collar at position x and time t. The axial drilling pressure at this location, as a function of time, affects the dynamic behavior of the drill string. The second derivative of the drill collar deflection at position x is used to describe the change in the bending radius. The linear density of a drill collar represents the mass per unit length, which is related to the volume and mass of the material. The cross-sectional area of ​​the drill collar is used to calculate the overall mechanical response. The second time derivative of drill collar deflection indicates deflection acceleration and is related to changes in dynamic response. The rock reaction force generated when the drill string interacts with the rock mass ahead represents the force exerted by the external environment on the drill string assembly. Hydraulic forces, mainly derived from fluid mechanics models, are used to describe the hydraulic effects on the drilling tools during the drilling process.

[0045] Subsequently, in order to address the unknown and complex distributed forces... This invention creatively introduces a geomechanical stiffness matrix, transforming the parameters into a set with clear physical meaning. This is parameterized. This step approximates the rock reaction force as the product of the stiffness matrix and the drill collar deflection: ; In its formula, Indicates horizontal force. This represents vertical force; the two combined form rock. Total force at a point It is a stiffness matrix, where each element represents the stiffness in a different direction. Stiffness of horizontal displacement to horizontal force. Stiffness of vertical displacement to horizontal force Stiffness of horizontal displacement to vertical force Stiffness of vertical displacement with respect to vertical force. It is a displacement vector, contained in time. Horizontal displacement at time and vertical displacement Indicates the location and time At that time, the reaction force of the rock on the drill bit, It is a geomechanical stiffness matrix, containing all stiffness coefficients. It displays the rock's response to deflection deformation in different directions in matrix form. It is a vector that contains deflection deformation components.

[0046] The geomechanical stiffness matrix This refers to the parameters characterizing the geomechanical environment of the drill string assembly, as described in this invention. Next, the downhole microprocessor unit employs a recursive inverse solution algorithm to identify the geomechanical stiffness matrix in real time. Preferably, the algorithm is the Extended Kalman Filter (EKF) algorithm. The downhole microprocessor unit processes the matrix... Each independent element (e.g.) Construct a state vector The EKF algorithm estimates the state vector by performing a two-step cycle of prediction and update. In the prediction step, the downhole microprocessor predicts the current state based on the state transition equation. This equation assumes that the geological environment has a certain degree of continuity, and its state change is a stochastic process: ; In its formula, and These are the estimated state vector values ​​for the current time step and the previous time step, respectively. This is process noise.

[0047] In the update step, this invention incorporates the geological feature information obtained in the preceding steps, namely the acoustic impedance distribution map, as a key prior constraint into the EKF algorithm. This fusion process can significantly improve the speed and accuracy of reverse identification.

[0048] Specifically, using measured deformation data Before updating the standard, the downhole microprocessor unit first interprets the acoustic impedance distribution map. If the map shows an area with extremely low acoustic impedance in a specific orientation, such as the right side (indicating the presence of a cavern), the downhole microprocessor unit will directly adjust the internal variables of the EKF algorithm.

[0049] For example, the adjustment includes: adjusting the state vector In the predicted values, the horizontal support stiffness The corresponding elements are forcibly corrected to a preset, near-zero minimum value. Furthermore, the process noise covariance matrix can also be adjusted. The numerical values ​​of the corresponding elements are used to reflect the certainty that geological parameters in that direction have undergone abrupt changes.

[0050] After the aforementioned prior correction based on geological feature information, the downhole microprocessor unit then executes the standard EKF update procedure. This step involves updating the measured deformation curve. The observed values ​​are compared with the theoretical deformation values ​​calculated forward using the corrected state vector and the aforementioned formula, and the resulting residuals are used as the final estimate of the corrected state vector.

[0051] Finally, through the aforementioned recursive, continuous computation incorporating prior information, the downhole microprocessor unit can output a parameter that is updated in real-time during drilling and accurately reflects the current geomechanical environment in front of the drill bit, namely the geomechanical stiffness matrix. This matrix is ​​immediately passed to the subsequent model prediction and control module as the physical basis for its predictions and decisions.

[0052] S3: Based on the aforementioned geomechanical environmental parameters, a model prediction control method is used to generate active guidance commands to guide the drilling direction; Specifically, the parameters characterizing the geomechanical environment, namely the geomechanical stiffness matrix, are identified in the aforementioned steps. Subsequently, the method of the present invention proceeds to the step of generating active guidance commands. This step is carried out and executed by the downhole microprocessor unit, and its core lies in applying the model predictive control (MPC) method to make forward-looking optimal decisions.

[0053] First, the model prediction and control module within the downhole microprocessor unit integrates the drill string dynamics model established in the previous steps with the geomechanical stiffness matrix identified in real time. Together, they form its internal prediction model. This prediction model can predict the drilling tool assembly's performance in a preset prediction time domain under a series of alternative control inputs in the future. The drilling trajectory within is denoted as Subsequently, the core task of the downhole microprocessor unit is to solve a multi-objective optimization problem, aiming to compute a set of optimal control sequences that minimize future overall costs. This optimization problem is solved by minimizing a comprehensive cost function. To achieve this, its mathematical expression is exemplarily as follows: ; In its formula, This represents summing the trajectory over a period of time, starting from the time step. arrive , This is the weighted squared error between the predicted output and the reference output. At time step The predicted output, At time step Reference output, A weighting matrix is ​​used to adjust the importance of different outputs. The weighted sum of squares of the unbiased function is used to control the penalty for the input. At time step The unbiased function, The weighting matrix is ​​used to adjust the penalty for the unbiased function. This term may represent a weight term related to the state error. The weight of this item At time step The state error, The squared penalty for controlling input changes. At time step The control input changes, A weighted matrix that penalizes changes in control input.

[0054] The cost function further includes a drill bit force balance term. This item is based on the geomechanical stiffness matrix identified in real time. The calculated imbalance of the net external force acting on the drill bit is used to penalize this imbalance. Introducing this factor guides the drill bit to find a more evenly distributed force path, resulting in reduced drill string vibration and improved borehole wall quality. For example, this imbalance force... It can be defined as being with The asymmetry of matrices is relevant, for example: ; In its formula, The square norm of the unbiased function represents the imbalance state of the system. Lateral stiffness, that is, stiffness under horizontal displacement. Longitudinal stiffness, i.e., stiffness under vertical displacement. Stiffness in the lateral direction relative to the longitudinal direction. Longitudinal stiffness relative to lateral response control The weight of the effect of the term in the total unbiased function. control The weight of the influence of the term in the total unbiased function.

[0055] The cost function further includes a rock-breaking energy term. This parameter is used to penalize high-energy-consuming drilling methods, guiding the drill bit to actively avoid hard rock and find easier drilling paths, thus improving overall drilling efficiency. Rock-breaking energy Based on drilling parameters and characteristics of rock strength The matrix is ​​estimated.

[0056] For example, rock-breaking energy It can be estimated by the following formula: ; In its formula, Indicates a proportional relationship or is proportional to... For real-time drilling pressure, For real-time torque, For drilling rate, This is the conversion factor for torque to energy consumption. for The trace of a matrix This is the weighting coefficient of this term in the total cost function.

[0057] In addition, the cost function may also include a control increment term. This item is used for penalty control commands. rate of change This is to ensure smooth operation of the active guide actuator and avoid frequent and violent movements of the actuator. This is the corresponding weight matrix.

[0058] During the optimization process, the downhole microprocessor unit also needs to consider physical constraints, such as the maximum output thrust of the active guide actuator, i.e. Finally, the downhole microprocessor unit employs efficient online optimization algorithms, such as quadratic programming (QP) or nonlinear programming (NLP) solvers, to calculate the cost function while satisfying all constraints. Minimize the optimal control sequence .

[0059] Based on the rolling time-domain execution strategy of model predictive control, the downhole microprocessor unit extracts only the first control command from the optimal sequence. This is then used as the final active guidance command generated at the current moment to guide the drilling direction. This command is immediately output to the controller of the active guidance actuator to drive its physical actions.

[0060] S4: Drive the active guidance actuator in the drill string assembly to execute the active guidance command; Specifically, following the aforementioned steps, the downhole microprocessor unit generates the optimal active guidance command based on the model predictive control method. Subsequently, the method of this invention drives the active guidance actuator to translate the instruction into precise physical intervention on the drilling trajectory. This step is a key link connecting intelligent decision-making with the physical world and realizing closed-loop control.

[0061] The active guidance command This is a digitized control vector whose dimensions correspond to the number of independently actuable push elements in the active guide actuator. For example, for a system containing three push blocks, this instruction could be represented as... Each element defines the target output thrust value of the corresponding push block, in Newtons. This instruction is sent from the downhole microprocessor unit to the dedicated embedded controller of the active guidance actuator via an internal high-speed data bus. This controller is an independent microprocessor unit responsible for performing low-order, high-precision servo control tasks.

[0062] The controller of the active guidance actuator receives instructions. Then, a safety check is performed first to confirm that the command value is within the preset safety range. Subsequently, the controller starts an independent closed-loop force control subroutine for each pushing element that needs to be activated. For example, this subroutine can use a proportional-integral-derivative (PID) control algorithm.

[0063] The active guiding actuator specifically includes a miniature hydraulic pump driven by a brushless DC motor, an accumulator for stabilizing system pressure, and high-speed electromagnetic proportional valves corresponding to each push block. To achieve closed-loop force control, a pressure sensor or strain gauge-type force sensor is further installed on the hydraulic cylinder of each push block or its supporting structure to monitor the actual applied thrust in real time. The controller's PID algorithm will determine the target thrust of the command. (Right now (elements in the data) and the actual thrust fed back by the sensor Compare and calculate the error The controller generates a control voltage based on this error. ; In its formula, The control signal at time t is used to adjust the system's behavior to achieve the target. This represents the control error at time t. The proportional part indicates that the control signal is proportional to the current error. Proportional gain is used to adjust the sensitivity of the response. The integral part represents the cumulative error over time. Integral gain affects the system's ability to compensate for long-term errors. The integral of the error from 0 to t represents the accumulation of historical errors. The differential part represents the response to the current rate of change of error. Differential gain is used to adjust the system's sensitivity to changes in error. Indicates error The derivative with respect to time describes the rate at which the error changes.

[0064] The controller will calculate the control voltage. The voltage is applied to the drive coil of the designated electromagnetic proportional valve. This voltage precisely controls the displacement of the valve core, thereby precisely regulating the flow and pressure of high-pressure hydraulic oil from the main hydraulic channel to the corresponding piston chamber of the push block.

[0065] The piston is under controlled hydraulic pressure Under the action of the drill bit, it moves and pushes the push block, made of high-strength wear-resistant material (such as one inlaid with hard alloy), out of the drill string body. The actual thrust generated by the push block against the wellbore is... With hydraulic pressure and piston area Directly related, that is The actual thrust is continuously monitored by the force sensor, forming a tight low-order control closed loop. This ensures that the actual output force accurately and quickly tracks the command target force, unaffected by environmental changes such as downhole temperature, pressure, and rock wall hardness.

[0066] The precise lateral force applied by the pusher block to the wellbore generates an equal and opposite lateral force on the drill string assembly, according to Newton's third law. This lateral force exerts a bending moment on the entire drill string assembly behind the drill bit, causing the drill bit to tilt slightly. As the entire drill string assembly rotates, the tilted drill bit preferentially cuts the rock in its tilt direction, thus deflecting the borehole trajectory in the desired direction, achieving the pusher-guided principle.

[0067] The execution of this physical intervention is a crucial step in the entire control process. The result—the change in the deformation state and spatial attitude of the drill string assembly—is captured instantly by the sensing module in this invention. This instantaneous change in physical state serves as a new initial condition, inputting it into step S1 of the next control cycle, thereby initiating a new round of rolling optimization processes of perception-identification-decision-execution. This high-frequency, seamlessly connected closed-loop operation enables the method of this invention to achieve continuous, proactive, and forward-looking control of the drilling process.

[0068] The above-described embodiments are merely illustrative of the present invention. Any equivalent embodiments made by those skilled in the art, without departing from the scope of the technical features disclosed in the present invention, using partial modifications or alterations to the technical content disclosed in the present invention, shall still fall within the scope of the technical features of the present invention.

Claims

1. A method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas, characterized in that, Includes the following steps: S1: Obtain real-time deformation data of the drill string assembly during the drilling process, and detect the geological features of the front end of the drill string assembly; S2: Based on the real-time deformation data and the geological feature information, the parameters characterizing the geomechanical environment of the drill string assembly are identified in reverse through the preset drill string dynamics model; S3: Based on the aforementioned geomechanical environmental parameters, a model prediction control method is used to generate active guidance commands to guide the drilling direction; S4: Drive the active guidance actuator in the drill string assembly to execute the active guidance command.

2. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 1, characterized in that, The steps for acquiring real-time deformation data include: the drill string assembly is equipped with a downhole microprocessor unit and a distributed strain sensor array, the downhole microprocessor unit has a built-in fiber grating demodulation module, and the distributed strain sensor array includes FBG sensors; During drilling with the drill string assembly, the fiber optic grating demodulation module emits broadband light to the FBG sensor at a preset high frequency and monitors its reflection spectrum in real time. When the drill collar of the drill string bends or twists due to force, the central reflection wavelength of the FBG sensor shifts, and the fiber optic grating demodulation module collects this wavelength shift. The downhole microprocessor unit converts the fiber optic grating into axial strain at the corresponding location based on the sensing principle of the fiber optic grating. ; Subsequently, the axial position of the drill collar was calculated using the downhole microprocessor unit. and time The curvature of the bending curvature This relationship is expressed by the following formula: ； In its formula, This refers to the axial strain at the corresponding position on the drill collar surface, calculated based on the wavelength drift. The radial distance from the neutral axis of the drill collar to the fiber core of the FBG sensor; The downhole microprocessor unit uses a numerical integration algorithm to process the curvature function. Along the axial direction A first integration yields the slope of the drill collar's deflection curve; a second integration reconstructs the real-time deformation curve of the entire sensor drill collar in three-dimensional space, which is defined as... This curve accurately describes the actual bending shape of the drill string at the current moment.

3. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 1, characterized in that, The steps for detecting geological features include: the drilling assembly is equipped with a downhole microprocessor unit and an acoustic transducer array; The downhole microprocessor unit drives the acoustic transducer array to actively emit controlled acoustic pulses towards the rock mass in front of the drill string at a preset pulse repetition frequency. The acoustic pulses propagate in the drilling fluid and reach the rock mass interface in front. Part of the energy is reflected to form an echo signal, which is captured by the same acoustic transducer array or a dedicated receiving transducer. The downhole microprocessor unit collects the echo signal and analyzes it. By calculating the flight time from emission to reception of the echo signal, the distance of the geological discontinuity interface in front is determined. Simultaneously, the downhole microprocessor unit analyzes the amplitude of the echo signal and calculates the acoustic reflection coefficient at the medium interface according to the following formula. : ； In its formula, The amplitude of the received echo signal. The amplitude of the original transmitted signal. Let be the acoustic impedance of the drilling fluid, and be a known or measurable parameter. The acoustic impedance of the rock mass to be probed ahead; By solving the above equation, the downhole microprocessor unit can infer the acoustic impedance of the rock mass ahead. By scanning and integrating the detection results of all acoustic transducers, a qualitative acoustic impedance distribution map characterizing the area in front of the drill bit can be generated. This acoustic impedance distribution map intuitively reveals geological features: areas with acoustic impedance values ​​close to zero indicate the presence of caves or large open fractures; areas with dramatic step changes in acoustic impedance values ​​indicate the presence of soft and hard rock layer interfaces.

4. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 1, characterized in that, The geomechanical environmental parameters are a geomechanical stiffness matrix, and each element of the matrix represents the support stiffness and coupling effect of the rock mass surrounding the drill bit in different directions.

5. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 1, characterized in that: The preset drill bit dynamics model is mathematically expressed as a partial differential equation: ； In its formula, The elastic modulus of the drill collar reflects the stiffness properties of the material. Let be the moment of inertia of the drill collar section, and represent the effect of the material shape on its bending capacity. This represents the fourth derivative of the drill collar at position x and time t. The axial drilling pressure at this location, as a function of time, influences the dynamic behavior of the drill string. The second derivative of the drill collar deflection at position x is used to describe the change in the bending radius. The linear density of the drill collar represents the mass per unit length. This represents the cross-sectional area of ​​the drill collar, used to calculate the overall mechanical response. The second time derivative of the drill collar deflection indicates the deflection acceleration and is related to the change in dynamic response. This represents the rock reaction force generated when the drill string interacts with the rock mass ahead, indicating the force exerted by the external environment on the drill string assembly. Hydraulic force, used to describe the hydraulic impact on the drilling tool during drilling; Subsequently, the geomechanical stiffness matrix was introduced. Parameterizing it, the geomechanical stiffness matrix These are the parameters of the aforementioned geomechanical environment, and the formula is as follows: ； In its formula, Indicates horizontal force. This represents vertical force; the two combined form rock. Total force at a point It is a stiffness matrix, where each element represents the stiffness in a different direction. Stiffness of horizontal displacement to horizontal force. Stiffness of vertical displacement to horizontal force Stiffness of horizontal displacement to vertical force Stiffness of vertical displacement with respect to vertical force. It is a displacement vector, contained in time. Horizontal displacement at time and vertical displacement Indicates the location and time At that time, the reaction force of the rock on the drill bit, It is a geomechanical stiffness matrix, containing all stiffness coefficients. It displays the rock's response to deflection deformation in different directions in matrix form. It is a vector that contains deflection deformation components.

6. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 5, characterized in that, The reverse identification steps include: using the Extended Kalman Filter (EKF) algorithm to transform the matrix... Each independent element is used to construct a state vector. The EKF algorithm estimates the state vector by performing two iterative steps: prediction and update. In the prediction step, the state at the current moment is predicted according to the state transition equation, which assumes that the geological environment has a certain continuity and that its state change is a stochastic process. ； In its formula, and These are the estimated state vector values ​​for the current time step and the previous time step, respectively. For process noise: In the update step, the geological feature information obtained in the aforementioned steps is incorporated into the EKF algorithm as a key prior constraint.

7. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 1, characterized in that: The model predictive control method employs a cost function, which includes at least one trajectory tracking term used to penalize the geometric deviation between the predicted trajectory and the design trajectory.

8. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 7, characterized in that, The cost function also includes a drill bit force balance term, calculated based on the geomechanical stiffness matrix, used to penalize drill bit force imbalance.

9. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 7, characterized in that, The cost function also includes a rock-breaking energy term, estimated based on the geomechanical stiffness matrix and drilling parameters, used to penalize high energy consumption.

10. The method for grouting and water plugging drilling correction in limestone sections of subway tunnels in karst areas according to claim 1, characterized in that: The active guiding actuator includes a miniature hydraulic pump driven by a brushless DC motor, an accumulator for stabilizing system pressure, multiple push blocks, and a high-speed electromagnetic proportional valve corresponding to each push block.

Images